116 commits

Author	SHA1	Message	Date
Shikhar Kumar	94d29e154e	Temporary fix to address precision discrepancies with weightfactors numpy array	2019-03-05 18:57:58 -05:00
Paul Romano	368f89697d	Use Position and Direction in Particle class	2019-03-01 16:00:36 -06:00
Paul Romano	d790cdfb46	Update list of publications, fix documentation for CMFD	2019-01-30 22:15:48 -06:00
Shikhar Kumar	9f63b5a3b7	Use kwargs instead of kwargs.keys()	2018-12-20 00:40:08 -05:00
Shikhar Kumar	bb45f220d4	clean up CMFD testing harness, adjoint logic, use run_in_memory	2018-12-19 02:04:11 -05:00
Shikhar Kumar	b87515b7fb	Update output string, kwargs, docstring	2018-12-14 20:33:13 -05:00
Shikhar Kumar	d4cfb81114	Fix test suite, remove cmfd from openmc/__init.py	2018-12-13 17:17:59 -05:00
Shikhar Kumar	5c04eb0d3b	Update test suite with 2g, ng solvers	2018-12-12 12:20:48 -05:00
Shikhar Kumar	2874b4a2f9	Address @paulromano and @smharper comments	2018-12-11 18:00:06 -05:00
Shikhar Kumar	ee9a6ab065	So begins the CMFD Fortran purge :D (~3.6K lines)	2018-11-24 00:13:37 -05:00
Shikhar Kumar	ab2f3bfc75	Remove keff_temp definition	2018-11-12 17:44:57 -05:00
Shikhar Kumar	e149c68a0e	Remove unused solvers, rely solely on vectorized numpy and c++ linear solver	2018-11-12 17:19:18 -05:00
Shikhar Kumar	e2c2a03220	Move solver to C++	2018-11-12 17:15:50 -05:00
Shikhar Kumar	70cced45bb	Some more minor fixes	2018-11-12 17:10:42 -05:00
Shikhar Kumar	cb932673c5	Minor changes to storing boolean array of is_adj_ref directly for each direction	2018-11-12 17:10:18 -05:00
Shikhar Kumar	58fdec12a1	Get rid of old vectorized functions, use updated versions	2018-11-12 17:09:45 -05:00
Shikhar Kumar	37030fed2f	Split between vectorized and updated-vectorized versions of CMFD	2018-11-12 17:06:27 -05:00
Shikhar Kumar	3f0878dd36	Precompute array indices, clean up readability more, fix bug with coremap axis swapping	2018-11-12 17:00:43 -05:00
Shikhar Kumar	d57592a031	Initial attempt at making vectorized functions more readable	2018-11-12 17:00:27 -05:00
Shikhar Kumar	b4d7ddbd48	Switch index of current tally direction and num groups to be consistent with openmc tallies	2018-11-12 16:56:52 -05:00
Shikhar Kumar	4a03fc9a03	Add more flushes for print statements	2018-10-17 11:45:28 -04:00
Shikhar Kumar	e0363898b7	Correct surface currents tally on Python side, throw error if not in CE mode	2018-10-17 11:27:57 -04:00
Shikhar Kumar	84b0614128	Merge branch 'develop' into cmfd-capi	2018-10-17 11:23:06 -04:00
Shikhar Kumar	2a709a0a56	Remove subdivision of next_batch, CAPI CMFD no longer preserves order of calls from Fortran side	2018-09-20 10:03:08 -04:00
Shikhar Kumar	91180f08a7	minor changes that should also be cherry-picked to shikhar413::cmfd-capi	2018-09-19 20:24:50 -04:00
Shikhar Kumar	1633d5a749	Bug fixes for 1grp-{1d,2d,3d} cases	2018-09-12 14:37:08 -04:00
Shikhar Kumar	e8176efa36	Remove functionality to handle restart batches, add comments for all CMFD instance variables and methods	2018-09-06 19:42:02 -04:00
Shikhar Kumar	7246266693	Solve for general hxyz, output cmfd timing stats	2018-08-30 14:08:00 -04:00
Shikhar Kumar	c5477e66f6	Cleanup code, write matrices/vector to file from Python, write cmfd output, populate matrices directly as csr format	2018-08-30 02:30:51 -04:00
Shikhar Kumar	84dad25c3d	Vecotrize source reweighting, merge vectorized and non-vectorized functions (compute_dhat, dtilde, build_matrix)	2018-08-29 00:21:38 -04:00
Shikhar Kumar	269e71f9d7	Fix travis issue, mpi4py not defined in some test cases	2018-08-28 16:19:51 -04:00
Shikhar Kumar	3631d15f51	Say hello to a new CMFD solver (Finish particle reweight)	2018-08-28 14:11:34 -04:00
Shikhar Kumar	5d31649e6d	Include vectorized versions of build matrices	2018-08-27 03:42:43 -04:00
Shikhar Kumar	c2858bc941	Clean up comments, add vector_write function in Fortran, create vectorized calc_fission_source	2018-08-26 22:45:58 -04:00
Shikhar Kumar	961178ed80	Vectorized version of compute_dtilde	2018-08-24 23:20:03 -04:00
Shikhar Kumar	1fde7f8599	Add vectorized version of compute_dhat	2018-08-24 15:38:49 -04:00
Shikhar Kumar	07f2b3ac3d	Implement CMFD solver, omp_num_threads now command line arg for CMFDRun.run()	2018-08-23 23:35:29 -04:00
Shikhar Kumar	e14826f7ff	Build production and loss matrices (not vectorized), print matrices at each batch in Fortran	2018-08-22 23:42:53 -04:00
Shikhar Kumar	d55fc1e9ce	Replicate cmfd_data.F90 with Python, still need to vectorize compute_dhat and compute_dtilde	2018-08-22 02:05:41 -04:00
Paul Romano	fbf7713a33	Split up Geometry.from_xml, rename clean_xml -> _xml	2018-08-21 07:47:29 -05:00
Shikhar Kumar	6798cafde1	Finish implementing compute_xs	2018-08-20 17:56:22 -04:00
Shikhar Kumar	2f43fa3960	Set cmfd tally type and estimator through C API	2018-08-16 20:12:04 -04:00
Shikhar Kumar	3a73450ed9	Replicate all of cmfd_header.F90 and cmfd_input.F90 through CAPI	2018-08-16 02:23:20 -04:00
Shikhar Kumar	6080c3a911	Create skeleton for cmfd execute through CAPI	2018-08-15 12:23:48 -04:00
Shikhar Kumar	977dec7a5a	Remove cmfd.py notes to separate file	2018-08-14 23:33:59 -04:00
Shikhar Kumar	cfbf0af96c	Incorporate comments from ANL meeting, remove reset property from tally_header.F90	2018-08-08 23:15:27 -04:00
Shikhar Kumar	651fb9c8f5	Get rid of unnecessary comments, have CMFDRun.run() run openmc completely without actually running CMFD	2018-07-31 23:38:10 -04:00
Shikhar Kumar	55761214fa	First pass at implementing CMFD through CAPI, for ANL meeting; most of cmfd_input.F90 replicated	2018-07-31 02:00:47 -04:00
Paul Romano	d3d6dc87dc	Remove a few more __future__ imports. Use collections.abc	2018-02-05 23:14:30 -05:00
Paul Romano	c428cee667	Remove __future__ and six imports	2018-02-05 23:14:30 -05:00